TMLS Model-Based Control of the Multimode Terrestrial-Aerial Wheeled-Bipedal Robot

In recent years, terrestrial-aerial hybrid locomotion robots have attracted significant research interest due to their enhanced energy efficiency and environmental adaptability. This article presents algorithmic optimizations for TLR, a novel terrestrial-aerial wheeled-bipedal robot (WBR). These optimizations enable the excellent ground mobility and novel modes such as standing long jumps. Additional modes expand the range of traversable terrains and enhance the robot’s flexibility and energy efficiency in complex tasks. A three-mass lift-spring (TMLS) model is proposed as a virtual model to capture the key dynamic characteristics of the TLR. By tuning the model parameters, leg-ground adaptability is achieved. A ground contact detection algorithm based on the TMLS model is introduced for the first time in WBRs, enabling functionalities such as single-wheel locomotion and two-wheel vertical takeoff and landing. Compared to the commonly used linear wheel-spring-loaded inverted pendulum model, the TMLS model increases the maximum jump height by at least 2.5 times. With the assistance of the rotors, the pioneering standing long jump is realized. A hierarchical control architecture is proposed to integrate the heterogeneous multimode platform, reducing control complexity while ensuring orderly execution and smooth transitions across locomotion modes. Simulations and experimental results validate the proposed algorithms and control framework.